Difference between double value in Oracle and Java (IEEE 754)

Question

BINARY_DOUBLE in Oracle and Double in Java use IEEE 754 standard.

But there is the difference in their accuracy.

For example value:

456.67d

Oracle:

declare
  a BINARY_DOUBLE := 456.67d;
begin
  SYS.DBMS_OUTPUT.PUT_Line(TO_CHAR(a,'9.99999999999999999999EEEE'));
end;

Result: 4.56670000000000020000E+02

Java:

Double d = 456.67d;
DecimalFormat formatter = new DecimalFormat("0.000000000000000000000E000");
System.out.println(formatter.format(d));

Result: 4.566700000000000000000E002

The value in Java is not as accurate as in Oracle.

Online converters said that the most accurate representation for 456.67d is:

4.56670000000000015916157281026E2

So why in Java accuracy is not the same like in Oracle? And how I can get in Java more accurate value?

Answer 1

Use BigDecimal:

Double d = 456.67d;
BigDecimal bd = new BigDecimal( d );
bd.setScale( 50 );
System.out.println( bd );
DecimalFormat formatter = new DecimalFormat("0.000000000000000000000E000");
System.out.println( formatter.format( bd ) );

Output:

456.67000000000001591615728102624416351318359375
4.566700000000000159162E002